Antipsychotic prescription to identify delirium: results from two cohorts

OBJECTIVES: Detection of delirium in hospitalized patients remains challenging. The objective was to determine if the prescription of antipsychotic medications was associated with delirium. PATIENTS AND METHODS: Two patient cohorts were utilized from a tertiary Veterans Affairs hospital: a palliative care retrospective cohort and a prospective medical cohort. Patients prescribed outpatient antipsychotics were excluded. Retrospectively, delirium was identified using a validated medical record-review instrument. Prospectively, a clinical expert assessed patients for delirium daily using a standardized interview. Acute antipsychotic medication administration was recorded from the electronic medical record. RESULTS: In the retrospective cohort (n=217), delirium was found in 31% (n=67) and antipsychotic use in 18% (n=40) of patients. Acute antipsychotic use indicated delirium with 54% sensitivity and 97% specificity. In the prospective cohort (n=100), delirium developed in 23% (n=23) and antipsychotics were used in 5% (n=5) of patients. The sensitivity and specificity of acute antipsychotic use was 22% and 100%, respectively. CONCLUSION: Hospitalized patients who are acutely prescribed antipsychotics are likely to have delirium, but not all patients with delirium will be identified with this method. In health systems, utilization of the prescription of acute antipsychotics can be an efficient and specific method to identify delirious patients for targeted intervention.